[PLUG] Please Sign The Petition

[Amarendra Godbole]

On Mon, Apr 13, 2015 at 4:56 PM, shirish शिरीष <shirishag75 at gmail.com> wrote:
[...]
> The idea that MS and Apple hit it right isn't also so true, whether
> for the so-called 'single user experience' or any other. Both MS and
> Apple had to spend millions of marketing dollars to sell their ribbon
> interfaces, UI's etc. There are reasons why Vista and Windows 8.1
> failed. In both those failures UI had lot to blame.
>
> What you seem to not note is that for most people, they are accustomed
> to the first GUI which came to them and obviously when you start with
> any instrument/machine however good or bad it is, you become 'trained'
> on it. There have been lot of studies done by neutral groups putting
> kids in a control group, and kids who were introduced  'MS-Windows' as
> well as kids who were shown a free software desktop ' GNOME' or 'KDE'
> . The understanding between kids who experienced MS-Windows and kids
> who were introduced to 'GNOME' and 'KDE' were at the same level with a
> little bit of advantage for the free desktop kids as they had much
> more to explore on the desktop than one is able to on an MS-Windows
> desktop (by itself.)

This exact argument was used 15 years ago -- about "familiarity" with
whatever comes first, when the MS-bashing was all the Linux camp could
dream about. Fast forward today, and this is the exact argument being
made, with no single recognizable Linux UI. I started my parents with
Ubuntu a few years ago, what happened? The UI radically and frequently
changed in the last 10 years, without any backward compatibility (yes
our home desktop could not run the later, power and graphics hungry
GNOME UIs' dished out by Ubuntu). My parents dumped Ubuntu, since it
was too much of a learning curve for them with every upgrade being
more "shiny". My argument is not so much with the familiarity, but
with the callousness popular Linux distros treat their end users with,
and still expect the masses to adopt their ever changing, power hungry
user interfaces.

Not a problem with Windows XP - over a decade the "base UI" that the
user is familiar with hasn't changed, irrespective of how many service
packs and patches were applied. The lack of a single unified user
interface that doesn't change for a decade is what masses want (we
aren't dealing with geeks you see).

>> Where are the replacement "right tools"?
>
> The replacement "right tools" are all here. What most of the tools are
> waiting for are users. If you have not got the 'right tool' it is
> possible it is not developed, it's also possible it's not known or
> marketed and hence is not there. Where 'free software' lacks is not
> that we don't have tools, we have more tools than we know what we do,
> what we lack is more often than not 'visibility'. I could share quite
> a number of tools/apps. who have gone that way and are going to the
> software grave but will share one just so people don't think I'm
> talking out of my ass.
>
> There is this pretty nice tool/utility which came out of the GNOME
> table sometime back called 'gnome-web-photo' .  As the short
> description goes 'Create snapshot images and print web pages from the
> command line' or the long description 'GNOME Web Photographer is a
> tool to generate full-size image files and thumbnails from HTML files
> and web pages. It can also be used to print  those.'
>
> Now AFAIK there isn't an equivalent to this tool in the free software
> world. I used to simply say the name of the application use one or
> more options given in the --help option and have a screenshot of a
> page which scrolled once or twice vertically which by conventional
> means would make me take couple of snapshots and then in some tool
> join the 2-3 pieces and cut out any over-lapping images. This made it
> so easy. Unfortunately, the tool was never marketed properly and it
> seems it will go to the grave soon as the code hasn't been touched in
> over a year and the developer is not responding to the bug-reports and
> I have been suffering because the tool does not wok anymore because
> the underlying GTK library has chaned. Hence, I do grant the point
> though that we are weak in marketing but not in development, or
> anything else for that matter.
>
> As far as the points being toted out in favor of 'unified strategy'
> are concerned, I think the successes we have had so far would have
> come if we had 'unified strategy'. Git wouldn't even have been born if
> we didn't have the 'let a thousand flowers bloom' strategy. While a
> corporation can make or attempt to have 'unified strategies'
> communities are better/stronger  when they have freedom to explore,
> criticize etc. If everybody is to be married to a unified strategy,
> then democracy doesn't have a place in society, as democracy gives
> everybody an unalienable right to believe in whatever they ought to.
> It is precisely that there isn't a unified strategy that every 8 weeks
> we have a new kernel. If we had unified strategy that would have been
> 4 years.

I believe the lack of unified interface (maybe a "dumb mode" switch)
would have gone a long way in accelerating the adoption. I never
argued against being on the bleeding edge, and allowing people to
tinker and improve -- but these are only the 20% of geeks and nerds.
The rest 80% want a dumb mode, and they don't care about anything
else.

Linux distros' are "unified inside, diverse outside" - good for geeks
and nerds. There should be "diverse inside, unified outside" for
masses. Maybe that'd pave way to more kernel innovation

Apple's user interface is an engineering/technical feat, irrespective
of their marketing. Can you say the same for GNOME or KDE in terms of
reliablity, stability as was as familiarity of look and feel? Heck, my
Microsoft Word on Mac is rock solid stable and hasn't crashed on me
once, while LibreOffice fails left right and center. Microsoft Office
for Mac is a ground up designed and built, and not a port. I don't see
that stability from any of the Linux-based word processing, or email
tool. Why?

My requirement for mass adoption would be like these (the "mass mode"
of Linux, geeks will often term it as "dumb mode"):
1. No changes to the user interface for a decade. Next change in 2025.
2. Explosive innovation to the file system, the kernel, and everything
that isn't visible. The nuts and bolts.
3. One tool for one job: spreadsheet, word editor, presentation, photo
editor, music player, video player, database, browser, email client,
and other common usage utilities (maybe two dozen).
4. One banner, no distro bloat.
5. "Kid mode" with restrictions on certain utilities, intelligently
defined by default.
6. "Geek mode" that unleashes full power of the OS, in and out - that
also allows UI customizations, as well as tinkering.

I will focus on polishing and polishing #3 till they are rock solid
stable, and secure. Feature bloat isn't desired, but stability and
reliability is a MUST.

> What we should be asking is do we want to become a MS-Windows clone or
> an Apple's clone ?
> Both of them have made lots of money but by doing deals with
> Governments and all sorts of nefarious activities that they have
> created.

No, setting priorities right would do magic. How about focusing on
uniformity, reliability, stability, correctness of the existing
offering, than feature bloat and next-gen systems? I am surprised the
lessons of the command-line world did not translate right to
UI-world...

>> - How do you propose to replace the extensive supply-chains, and other
>> public utility systems that are deployed on Windows?
>> - How do you propose to replace the extensive manufacturing systems
>> that are deployed on Windows?
>> - How do you propose to replace the extensive life-critical software
>> in hospitals that is deployed on Windows?
>
> Are you saying that these systems will be in stati ? All of the
> examples you have shared are pretty much dynamic software systems and
> they have to be changed. In fact, if you look at all the examples
> quoted all of them are pretty much on the embedded side of things and
> that's a place where GNU/Linux is situated rather well.

Certainly, it will be a welcome change, when my LibreOffice will crash
once in a month, than once a day.

> Apart from that, it will also benefit to have some or many of such
> systems on a different OS from a security perspective as well. It is a
> known and proven fact (time and again) that homogeneous systems are
> more easily to be cracked/brute-forced then systems in a heterogeneous
> environment.

And it is also a known fact that less lines of code are more secure
than more. OpenBSD ls has ~1400 loc as against ~5000 of GNU ls (a
straight up comparison isn't feasible, since GNU ls is heavily
commented, while the OpenBSD is not. However, this is just one
utility...).

Linux never had the focus on security, it merely "happened". Also,
there isn't a point, since Linux happily accepts "binary blobs"
rendering any and all security effectively useless. Hardening Linux is
like debating the strength of a padlock on a tent. ;-)

Security happens to be my bread and butter, in addition to primary
interest, and I do not wish to drag this thread in that direction.
Let's stop here for security. :-)

> I don't think it is as much as 'failures' which has been the cause of
> customer irritation/annoyance it's more a lack of number of friendly
> Linux geeks as well as paid support in case things go wrong.

That statement itself admits the difficulty of providing support for
Linux - no other commercial operating system requires "geeks" for
support. ;-)

[...]

> While I do agree with the part of the BSD license being a superior
> license in the sense it doesn't 'infect' other people's work as the
> GNU license does, it also leads me to wonder why is it that most BSD
> distributions are not well-maintained (as in frequent releases, new
> developments etc.) There are only 2-3 major innovations I can recall
> from the BSD stable, that is the BSD real-time kernel (we have almost
> feature parity there), ZFS ( BTRFS, sort of when it's ready) and the
> BSD Jails (we do have partial working implementations but no clear
> answer yet.) apart from that haven't heard anything interesting in
> years. I would welcome if you were able to share more on their behalf.

You are misinformed. Actually, FreeBSD (latest release Oct 2014),
OpenBSD (latest release Nov 2014), NetBSD (latest release Oct 2014)
and DragonFlyBSD (latest release Nov 2014) are very well maintained -
OpenBSD has a predictable release schedule, every 1st of May and Nov.
If you ever want to look at quality and clean code, look no further
than any of the *BSD. Most C programmers can attest to this fact - ask
them which code is clean and easy to follow of the *BSD or Linux, and
they will tell you their flavor of BSD.

Most people do not hear the *BSD, since they focus on a subset of
problems and try to solve them in a better fashion. Unlike Linux,
which happens to stick its fingers in all pies, and ends up with
subpar code.

But again, *BSD is for another day, another discussion.

[...]

Again, thanks for a balanced perspective. Hopefully we meet in person
when this discussion can continue.

-Amarendra